Summary:
This diff adds support for concurrent adds to the skiplist memtable
implementations. Memory allocation is made thread-safe by the addition of
a spinlock, with small per-core buffers to avoid contention. Concurrent
memtable writes are made via an additional method and don't impose a
performance overhead on the non-concurrent case, so parallelism can be
selected on a per-batch basis.
Write thread synchronization is an increasing bottleneck for higher levels
of concurrency, so this diff adds --enable_write_thread_adaptive_yield
(default off). This feature causes threads joining a write batch
group to spin for a short time (default 100 usec) using sched_yield,
rather than going to sleep on a mutex. If the timing of the yield calls
indicates that another thread has actually run during the yield then
spinning is avoided. This option improves performance for concurrent
situations even without parallel adds, although it has the potential to
increase CPU usage (and the heuristic adaptation is not yet mature).
Parallel writes are not currently compatible with
inplace updates, update callbacks, or delete filtering.
Enable it with --allow_concurrent_memtable_write (and
--enable_write_thread_adaptive_yield). Parallel memtable writes
are performance neutral when there is no actual parallelism, and in
my experiments (SSD server-class Linux and varying contention and key
sizes for fillrandom) they are always a performance win when there is
more than one thread.
Statistics are updated earlier in the write path, dropping the number
of DB mutex acquisitions from 2 to 1 for almost all cases.
This diff was motivated and inspired by Yahoo's cLSM work. It is more
conservative than cLSM: RocksDB's write batch group leader role is
preserved (along with all of the existing flush and write throttling
logic) and concurrent writers are blocked until all memtable insertions
have completed and the sequence number has been advanced, to preserve
linearizability.
My test config is "db_bench -benchmarks=fillrandom -threads=$T
-batch_size=1 -memtablerep=skip_list -value_size=100 --num=1000000/$T
-level0_slowdown_writes_trigger=9999 -level0_stop_writes_trigger=9999
-disable_auto_compactions --max_write_buffer_number=8
-max_background_flushes=8 --disable_wal --write_buffer_size=160000000
--block_size=16384 --allow_concurrent_memtable_write" on a two-socket
Xeon E5-2660 @ 2.2Ghz with lots of memory and an SSD hard drive. With 1
thread I get ~440Kops/sec. Peak performance for 1 socket (numactl
-N1) is slightly more than 1Mops/sec, at 16 threads. Peak performance
across both sockets happens at 30 threads, and is ~900Kops/sec, although
with fewer threads there is less performance loss when the system has
background work.
Test Plan:
1. concurrent stress tests for InlineSkipList and DynamicBloom
2. make clean; make check
3. make clean; DISABLE_JEMALLOC=1 make valgrind_check; valgrind db_bench
4. make clean; COMPILE_WITH_TSAN=1 make all check; db_bench
5. make clean; COMPILE_WITH_ASAN=1 make all check; db_bench
6. make clean; OPT=-DROCKSDB_LITE make check
7. verify no perf regressions when disabled
Reviewers: igor, sdong
Reviewed By: sdong
Subscribers: MarkCallaghan, IslamAbdelRahman, anthony, yhchiang, rven, sdong, guyg8, kradhakrishnan, dhruba
Differential Revision: https://reviews.facebook.net/D50589
This is an Env implementation that mirrors all storage-related methods on
two different backend Env's and verifies that they return the same
results (return status and read results). This is useful for implementing
a new Env and verifying its correctness.
Signed-off-by: Sage Weil <sage@redhat.com>
Summary:
This diff is 1/3 in a sequence that introduces a skip list optimized for
a key that is a freshly-allocated const char*. The diff is broken into
pieces to make it easier to review. This piece only introduces the new
type by copying the existing SkipList, with mechanical naming changes
and reformatting.
Test Plan: new unit test
Reviewers: igor, sdong
Reviewed By: sdong
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D51279
* conversion from 'size_t' to 'type', by add static_cast
Tested:
* by build solution on Windows, Linux locally,
* run tests
* build CI system successful
Summary:
Under a tailing workload, there were increased block cache
misses when a memtable was flushed because we were rebuilding iterators
in that case since the version set changed. This was exacerbated in the
case of iterate_upper_bound, since file iterators which were over the
iterate_upper_bound would have been deleted and are now brought back as
part of the Rebuild, only to be deleted again. We now renew the iterators
and only build iterators for files which are added and delete file
iterators for files which are deleted.
Refer to https://reviews.facebook.net/D50463 for previous version
Test Plan: DBTestTailingIterator.TailingIteratorTrimSeekToNext
Reviewers: anthony, IslamAbdelRahman, igor, tnovak, yhchiang, sdong
Reviewed By: sdong
Subscribers: yhchiang, march, dhruba, leveldb, lovro
Differential Revision: https://reviews.facebook.net/D50679
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
Summary:
This patch allows rocksdb to persist options into a file on
DB::Open, SetOptions, and Create / Drop ColumnFamily.
Options files are created under the same directory as the rocksdb
instance.
In addition, this patch also adds a fail_if_missing_options_file in DBOptions
that makes any function call return non-ok status when it is not able to
persist options properly.
// If true, then DB::Open / CreateColumnFamily / DropColumnFamily
// / SetOptions will fail if options file is not detected or properly
// persisted.
//
// DEFAULT: false
bool fail_if_missing_options_file;
Options file names are formatted as OPTIONS-<number>, and RocksDB
will always keep the latest two options files.
Test Plan:
Add options_file_test.
options_test
column_family_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48285
Summary:
Using a TLS random instance for skiplist makes it smaller
(useful for hash_skiplist_rep) and prepares skiplist for concurrent
adds. This diff also modifies the branching factor math to avoid an
unnecessary division.
This diff has the effect of changing the sequence of skip list node
height choices made by tests, so it has the potential to cause unit
test failures for tests that implicitly rely on the exact structure
of the skip list. Tests that try to exactly trigger a compaction are
likely suspects for this problem (these tests have always been brittle to
changes in the skiplist details). I've minimizes this risk by reseeding
the main thread's Random at the beginning of each test, increasing the
universal compaction size_ratio limit from 101% to 105% for some tests,
and verifying that the tests pass many times.
Test Plan: for i in `seq 0 9`; do make check; done
Reviewers: sdong, igor
Reviewed By: igor
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D50439
Enable C4307 'operator' : integral constant overflow
Longs and ints on Windows are 32-bit hence the overflow
Enable C4309 'conversion' : truncation of constant value
Enable C4512 'class' : assignment operator could not be generated
Enable C4701 Potentially uninitialized local variable 'name' used
Summary:
This patch introduces OptionsSanityCheckLevel internally to enable
sanity check rocksdb options.
Utilities API will be added in the follow-up diffs.
Test Plan: Added more tests in options_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49515
Summary:
This patch introduces utilities/memory, which currently includes
GetApproximateMemoryUsageByType that reports different types of
rocksdb memory usage given a list of input DBs.
The API also take care of the case where Cache could be shared
across multiple column families / multiple db instances.
Currently, it reports memory usage of memtable, table-readers
and cache.
Test Plan: utilities/memory/memory_test.cc
Reviewers: igor, anthony, IslamAbdelRahman, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49257
C4101 'identifier' : unreferenced local variable
C4189 'identifier' : local variable is initialized but not referenced
C4100 'identifier' : unreferenced formal parameter
C4296 'operator' : expression is always false
Summary:
The goal of this diff is to create a simple stress test with focus on catching:
* bugs in compaction/flush processes, especially the ones that cause assertion errors
* bugs in the code that deletes obsolete files
There are two parts of the test:
* write_stress, a binary that writes to the database
* write_stress_runner.py, a script that invokes and kills write_stress
Here are some interesting parts of write_stress:
* Runs with very high concurrency of compactions and flushes (32 threads total) and tries to create a huge amount of small files
* The keys written to the database are not uniformly distributed -- there is a 3-character prefix that mutates occasionally (in prefix mutator thread), in such a way that the first character mutates slower than second, which mutates slower than third character. That way, the compaction stress tests some interesting compaction features like trivial moves and bottommost level calculation
* There is a thread that creates an iterator, holds it for couple of seconds and then iterates over all keys. This is supposed to test RocksDB's abilities to keep the files alive when there are references to them.
* Some writes trigger WAL sync. This is stress testing our WAL sync code.
* At the end of the run, we make sure that we didn't leak any of the sst files
write_stress_runner.py changes the mode in which we run write_stress and also kills and restarts it. There are some interesting characteristics:
* At the beginning we divide the full test runtime into smaller parts -- shorter runtimes (couple of seconds) and longer runtimes (100, 1000) seconds
* The first time we run write_stress, we destroy the old DB. Every next time during the test, we use the same DB.
* We can run in kill mode or clean-restart mode. Kill mode kills the write_stress violently.
* We can run in mode where delete_obsolete_files_with_fullscan is true or false
* We can run with low_open_files mode turned on or off. When it's turned on, we configure table cache to only hold a couple of files -- that way we need to reopen files every time we access them.
Another goal was to create a stress test without a lot of parameters. So tools/write_stress_runner.py should only take one parameter -- runtime_sec and it should figure out everything else on its own.
In a separate diff, I'll add this new test to our nightly legocastle runs.
Test Plan:
The goal of this test was to retroactively catch the following bugs: D33045, D48201, D46899, D42399. I failed to reproduce D48201, but all others have been caught!
When i reverted https://reviews.facebook.net/D33045:
./write_stress --runtime_sec=200 --low_open_files_mode=true
Iterator statuts not OK: IO error: /fast-rocksdb-tmp/rocksdb_test/write_stress/089166.sst: No such file or directory
When i reverted https://reviews.facebook.net/D42399:
python tools/write_stress_runner.py --runtime_sec=5000
Running write_stress, will kill after 5 seconds: ./write_stress --runtime_sec=-1
Running write_stress, will kill after 2 seconds: ./write_stress --runtime_sec=-1 --destroy_db=false --delete_obsolete_files_with_fullscan=true
Running write_stress, will kill after 7 seconds: ./write_stress --runtime_sec=-1 --destroy_db=false
Running write_stress, will kill after 5 seconds: ./write_stress --runtime_sec=-1 --destroy_db=false
Running write_stress, will kill after 8 seconds: ./write_stress --runtime_sec=-1 --destroy_db=false --low_open_files_mode=true
Write to DB failed: IO error: /fast-rocksdb-tmp/rocksdb_test/write_stress/019250.sst: No such file or directory
ERROR: write_stress died with exitcode=-6
When i reverted https://reviews.facebook.net/D46899:
python tools/write_stress_runner.py --runtime_sec=1000
runtime: 1000
Going to execute write stress for [3, 3, 100, 3, 2, 100, 1, 788]
Running write_stress for 3 seconds: ./write_stress --runtime_sec=3 --low_open_files_mode=true
Running write_stress for 3 seconds: ./write_stress --runtime_sec=3 --destroy_db=false --delete_obsolete_files_with_fullscan=true
Running write_stress, will kill after 100 seconds: ./write_stress --runtime_sec=-1 --destroy_db=false --delete_obsolete_files_with_fullscan=true
write_stress: db/db_impl.cc:2070: void rocksdb::DBImpl::MarkLogsSynced(uint64_t, bool, const rocksdb::Status&): Assertion `log.getting_synced' failed.
ERROR: write_stress died with exitcode=-6
Reviewers: IslamAbdelRahman, yhchiang, rven, kradhakrishnan, sdong, anthony
Reviewed By: anthony
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D49533
Test code errors are currently blocking Windows Release builew
We do not want spend time building in Release what we can not run
We want to eliminate a source of most frequent errors when people
check-in test only code which can not be built in Release.
This feature will work only if you invoke msbuild against rocksdb.sln
Invoking it against ALL_BUILD target will attempt to build everything.
* add validation is git found
* add validation is .git folder exists in project (project zip archive download without .git folder)
* get head commit SHA if git found and .git folder exists
Tested:
* configure project by CMake 3.0.0 successfully (with and without git), with project zip archive (without .git folder) and with project cloned from github
* configure project by command: cmake -G "Visual Studio 12 Win64"
* build solution by Visual Studio
* manually validate that file utils/build_version.cc contains valid head revision value
Summary: In MyRocks, it is sometimes important to get propeties only for the subset of the database. This diff implements the API in RocksDB.
Test Plan: ran the GetPropertiesOfTablesInRange
Reviewers: rven, sdong
Reviewed By: sdong
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48651
Summary:
We are cleaning up dependencies.
This diff takes a first step at moving memtable files to their own
directory called memtable. In future diffs, we will move other memtable
files from db to memtable.
Test Plan: make check
Reviewers: sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D48915
Summary: As part of cleaning up dependencies for tech debt week, we are moving ldb and sst_dump tools from util to tools, since they are tools.
Test Plan: make check
Reviewers: sdong
Reviewed By: sdong
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48747
* add -C ${CMAKE_CURRENT_SOURCE_DIR} to git can found revision from current source directory
Tested:
* configure project by CMake 3.0.0 successfully
* configure project by command: cmake -G "Visual Studio 12 Win64"
* build solution by Visual Studio
* manually validate that file utils/build_version.cc contains valid head revision value
Summary: manual_compaction_test.cc incorrectly in util. Moved to db.
Test Plan: make check
Reviewers: sdong
Reviewed By: sdong
Subscribers: dhruba
Differential Revision: https://reviews.facebook.net/D48687
Summary:
As part of tech debt week, we are cleaning up dependencies.
This diff moves db_test_util.[h,cc] from util to db directory.
Test Plan: make check
Reviewers: igor, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D48543
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
Add an optimized build config switch for faster test runs
Change compiler options to introduce more opitmizations and be more inline with MS internal switches.
Make appveyor build to utilize all the avaiable cores on the VM (parallel)
Introduce new appveyor configuration for daily test runs as it would take too long
to run db_test after each checkin even in paralell.
With some exclusions we make it in 38 minutes. We currently fail to install ramdisk during the build.
Add a powershell script to faicilitate paralell run for db_test cases.
Summary:
Fix hex2String performance issues by removing sscanf dependency.
Also fixed some edge case handling (odd length, bad input).
Test Plan: Created a test file which called old and new implementation, and validated results are the same. I'll paste results in the phabricator diff.
Reviewers: igor, rven, anthony, IslamAbdelRahman, kradhakrishnan, yhchiang, sdong
Reviewed By: sdong
Subscribers: thatsafunnyname, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D46785
Summary:
Builder and CompactionJob share a lot of fairly complex code. This patch
refactors this code into a separate class, the CompactionIterator. Because the
shared code is fairly complex, this patch hopefully improves maintainability.
While there are is a lot of potential for further improvements, the patch is
intentionally pretty close to the original structure because the change is
already complex enough.
Test Plan: make clean all check && ./db_stress
Reviewers: rven, anthony, yhchiang, sdong, igor
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46197
Summary:
Prototype of API to allow MyRocks to override default Mutex/CondVar used by transactions with their own implementations. They would simply need to pass their own implementations of Mutex/CondVar to the templated TransactionDB::Open().
Default implementation of TransactionDBMutex/TransactionDBCondVar provided (but the code is not currently changed to use this).
Let me know if this API makes sense or if it should be changed
Test Plan: n/a
Reviewers: yhchiang, rven, igor, sdong, spetrunia
Reviewed By: spetrunia
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D43761
Summary: Provide a way to specify a detailed static error message for a Status without incurring a memcpy. Let me know what people think of this approach.
Test Plan: added simple test
Reviewers: igor, yhchiang, rven, sdong
Reviewed By: sdong
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D44259
Summary:
As I keep adding new features to transactions, I keep creating more duplicate code. This diff cleans this up by creating a base implementation class for Transaction and OptimisticTransaction to inherit from.
The code in TransactionBase.h/.cc is all just copied from elsewhere. The only entertaining part of this class worth looking at is the virtual TryLock method which allows OptimisticTransactions and Transactions to share the same common code for Put/Get/etc.
The rest of this diff is mostly red and easy on the eyes.
Test Plan: No functionality change. existing tests pass.
Reviewers: sdong, jkedgar, rven, igor
Reviewed By: igor
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D45135
Summary:
Initial implementation of Pessimistic Transactions. This diff contains the api changes discussed in D38913. This diff is pretty large, so let me know if people would prefer to meet up to discuss it.
MyRocks folks: please take a look at the API in include/rocksdb/utilities/transaction[_db].h and let me know if you have any issues.
Also, you'll notice a couple of TODOs in the implementation of RollbackToSavePoint(). After chatting with Siying, I'm going to send out a separate diff for an alternate implementation of this feature that implements the rollback inside of WriteBatch/WriteBatchWithIndex. We can then decide which route is preferable.
Next, I'm planning on doing some perf testing and then integrating this diff into MongoRocks for further testing.
Test Plan: Unit tests, db_bench parallel testing.
Reviewers: igor, rven, sdong, yhchiang, yoshinorim
Reviewed By: sdong
Subscribers: hermanlee4, maykov, spetrunia, leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D40869
Summary: util/delete_scheduler_impl.cc is missing CMakeLists.txt, which fails Windows build. Add it.
Test Plan: Build in both of Windows and Linux
Reviewers: anthony, rven, yhchiang, IslamAbdelRahman
Reviewed By: IslamAbdelRahman
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D43641
Summary:
Add two unit tests for SyncWAL(). One makes sure SyncWAL() doesn't block writes in the other thread. Another one makes sure SyncWAL() doesn't wait ongoing writes to finish before being executed.
Create a new test file db_wal_test and move two WAL related tests from db_test to here.
Test Plan: Run the new tests
Reviewers: IslamAbdelRahman, rven, kradhakrishnan, kolmike, tnovak, yhchiang
Reviewed By: yhchiang
Subscribers: leveldb, dhruba
Differential Revision: https://reviews.facebook.net/D43605
Summary:
This diff adds CompactOnDeletionCollector in utilities/table_properties_collectors,
which applies a sliding window to a sst file and mark this file as need-compaction
when it observe enough deletion entries within the consecutive keys covered by
the sliding window.
Test Plan: compact_on_deletion_collector_test
Reviewers: igor, anthony, IslamAbdelRahman, kradhakrishnan, yoshinorim, sdong
Reviewed By: sdong
Subscribers: maykov, dhruba
Differential Revision: https://reviews.facebook.net/D41175
Nathan Bronson